pipermail date handling

[Fehwalker-users]

I moved my first list to mailman 2.1 and tried to re-generate the pipermail
archives (just, ya know, to see how it'd work :), and got the following
traceback. This is a list that's been around for a long time (archives back
to 1995), and the .mbox file was generated from a majordomo/hypermail
archive when it was moved to mailman 2.0 a couple of years ago.

Here's a ls -l of the .mbox file:

-rw-rw-r-- 1 mailman mailman 17741891 Jan 1 10:36 bryans-list.mbox

Here's the traceback:

Updating HTML for article 467
Updating HTML for article 468
Updating HTML for article 469
Updating HTML for article 473
Updating HTML for article 472
Pickling archive state into
/home/mailman-2.1/archives/private/bryans-list/pipermail.pck
Traceback (most recent call last):
  File "bin/arch", line 187, in ?
    main()
  File "bin/arch", line 175, in main
    archiver.processUnixMailbox(fp, start, end)
  File "/home/mailman-2.1/Mailman/Archiver/pipermail.py", line
544, in processUnixMailbox
    m = mbox.next()
  File "/usr/local/lib/python2.2/mailbox.py", line 34, in next
    return self.factory(_Subfile(self.fp, start, stop))
  File "/home/mailman-2.1/Mailman/Mailbox.py", line 79, in
scrubber
    return mailbox.scrub(msg)
  File "/home/mailman-2.1/Mailman/Mailbox.py", line 99, in scrub
    return self._scrubber(self._mlist, msg)
  File "/home/mailman-2.1/Mailman/Handlers/Scrubber.py", line
132, in process
    dir = calculate_attachments_dir(mlist, msg, msgdata)
  File "/home/mailman-2.1/Mailman/Handlers/Scrubber.py", line 93,
in calculate_attachments_dir
    datedir = safe_strftime(fmt, now)
  File "/home/mailman-2.1/Mailman/Handlers/Scrubber.py", line 77,
in safe_strftime
    return time.strftime(fmt, floatsecs)
TypeError: argument must be 9-item sequence, not None

I'm guessing it's a header problem in one of the messages, but I'm not sure
what.

Thanks,

Bryan

[http://sourceforge.net/tracker/index.php?func=detail&aid=660733&group_id=103&atid=100103]

[Fehwalker-users (fehwalker-users)]

After some investigation, the following (admittedly invalid) date headers
cause tracebacks similar to the above. After reformatting them I'm able to
generate the archive.

Date: 25 Aug 95 18.00
Date: Thursday, 30 October 1997 3:02pm PT
Date: Mon, 29 Nov 1999 Pacific Standard Time

Note that these messages *didn't* cause bin/arch|pipermail in mailman
2.0.x to fail, though it probably didn't parse them properly. Perhaps
trapping the error and/or skipping these messages might be more useful?

It also appears that bin/arch is throwing all messages with dates it can't
figure out (after the above were removed) into the current day. An example
can be seen at
http://lists.samurai.com/pipermail/bryans-list/2003-January/thread.html .
I'm unsure if there's any way to better handle this, but just wanted to
note it - I can create a separate bug report if it's important.

Thanks,

Bryan

[Followme (followme)]

I can confirm this with 2.1.2. Is there a simple way to make arch guess
dates
better?
I would settle for arch using the date of the previous message, but I
don't
know wher to start digging in the code.

[Cem-users (cem-users)]

Not being a python programmer I'm not sure how difficult it
would be to do this. But would it be possible to code it so
that it just uses the envelope date instead or maybe if it
can't get validly formatted data for the date it uses
currently it would use the envelope date?

The majority of the email that the archives here are having
this problem with have a valid envelope date on them.